The present invention relates to an apparatus for, and the processing of semiconductor substrates. In particular, the invention relates to improving accuracy in the patterning of thin films during substrate processing via the deposition of antireflective coatings containing silicon (Si), nitrogen (N), and, optionally, oxygen (O) at low deposition rates, using nitrogen as a dilutant.
Since semiconductor devices were first introduced several decades ago, device geometries have decreased dramatically in size. During that time, integrated circuits have generally followed the two year/half-size rule (often called Moore's Law), meaning that the number of devices that will fit on a chip doubles every two years. Today's semiconductor fabrication plants routinely produce devices with feature sizes of 0.5 .mu.m or even 0.35 .mu.m, and tomorrow's plants will be producing devices with even smaller feature sizes.
A common step in the fabrication of such devices is the formation of a thin film on a substrate by the chemical reaction of gases. When patterning thin films, it is desirable that fluctuations in line width and other critical dimensions be minimized. Errors in these dimensions can result in variations in device characteristics or open-/short-circuited devices, thereby adversely affecting device yield. Thus, as feature sizes decrease, structures must be fabricated with greater accuracy. As a result, some manufacturers now require that variations in the dimensional accuracy of patterning operations be held to within 5% of the dimensions specified by the designer.
Thin films are often patterned by etching away portions of the deposited layer. To this end, modern substrate processing systems often employ photolithographic techniques in such patterning processes. Typically, these photolithographic techniques employ photoresist (PR) or other photosensitive material. In conventional photolithographic processing, photoresist is first deposited on a substrate. A photomask (also known simply as a mask) having transparent and opaque regions that embody the desired pattern is positioned over the photoresist. When the mask is exposed to radiant energy (e.g., light), the transparent portions permit the exposure of the photoresist in those regions, but not in the regions where the mask is opaque. The radiant energy causes a chemical reaction in exposed portions of the photoresist. A suitable chemical, chemical vapor or ion bombardment process is then used to selectively attack either the reacted or unreacted portions of the photoresist. This process is known as developing the photoresist. With the remaining photoresist acting as a mask, the underlying layer may then undergo further processing. For example, material may be deposited, the underlying layer may be etched or other processing carried out.
Modern photolithographic techniques often involve the use of equipment known as steppers, which are used to mask and expose photoresist layers. Steppers often use monochromatic (single-wavelength) radiant energy (e.g., monochromatic light), enabling them to produce the detailed patterns required in the fabrication of fine geometry devices. As a substrate is processed, however, the topology of the substrate's upper surface becomes progressively less planar. This uneven topology can cause reflection and refraction of the incident radiant energy, resulting in exposure of some of the photoresist beneath the opaque portions of the mask. As a result, this uneven surface topology can alter the patterns transferred by the photoresist layer, thereby altering critical dimensions of the structures fabricated.
Reflections from the underlying layer also may cause a phenomenon known as standing waves. When a photoresist layer is deposited on a reflective underlying layer and exposed to monochromatic radiant energy (e.g., deep ultraviolet (UV) light), standing waves may be produced within the photoresist layer. In such a situation, the reflected radiant energy interferes with the incident radiant energy and causes a periodic variation in intensity within the photoresist layer in the vertical direction. Standing-wave effects are usually more pronounced at the deep UV wavelengths used in modem steppers than at longer wavelengths because many commonly used materials are more reflective at deep UV wavelengths. The use of monochromatic light, as contrasted with polychromatic (e.g., white) light, also contributes to these effects because resonance is more easily induced in monochromatic light. The existence of standing waves in the photoresist layer during exposure causes roughness in the vertical walls formed when the photoresist layer is developed, which translates into variations in line widths, spacing, and other critical dimensions. To achieve the requisite dimensional accuracy, two approaches have been taken, both of which entail the use of another layer in addition to the photoresist layer.
The first approach uses a relatively thick organic film known as an antireflective coating (ARC), deposited beneath the photoresist, that absorbs incident radiant energy so that reflection and refraction of the incident radiant energy are minimized. A disadvantage of such organic films is that they require more process steps, and, being polymer-based, are difficult to etch.
A second approach helpful in achieving the necessary dimensional accuracy is the use of a dielectric antireflective coating (DARC), usually a thin layer of silicon oxynitride (SiO.sub.X N.sub.Y) or silicon nitride (SiN.sub.X). The optical characteristics of a DARC are chosen to minimize the effects of reflections occurring at interlayer interfaces during the photolithography process. The DARC's absorptive index (k) is such that the amount of radiant energy transmitted in either direction is minimized, thus attenuating both transmitted incident radiant energy and reflections thereof. The DARC's refractive index (n) is matched to that of the associated photoresist material in order to reduce refraction of the incident radiant energy.
Such films may be formed, for example, by the chemical reaction of gases, a process referred to as chemical vapor deposition (CVD). Thermal CVD processes supply reactive gases to the substrate surface where, induced by high temperatures, chemical reactions take place to produce the desired film. In contrast, plasma-enhanced CVD (PECVD) processes promote excitation and/or disassociation of the reactant gases by the application of radio frequency (RF) energy to a reaction zone proximate to the substrate's surface, thereby creating a plasma of highly reactive species. The high reactivity of the released species reduces the energy (i.e., temperature) required for a chemical reaction to take place. These relatively low temperatures foster a more stable process and are therefore preferable when depositing a DARC.
The creation of DARCs necessitates the reliable control of optical and physical film parameters. These parameters normally include the film's refractive index (n), absorptive index (k), and thickness (t). A film's refractive and absorptive indices may be controlled by controlling the film's composition. For example, in a silicon oxynitride DARC, these indices may be altered by adding nitrogen-containing process gases.
Thickness, control of which is important in the deposition of many SiN.sub.X and SiO.sub.X N.sub.Y films, is a particularly important factor in determining the optical qualities of a DARC. Because a DARC uses interference to minimize reflected radiant energy, a DARC must be deposited to a thickness that provides the proper phase-shift of the radiant energy reflected from the lower surface. This cancellation may be accomplished by ensuring that light reflected at the DARC's lower surface (the interface between the DARC and the underlying layer) is 180 (or 540 or another odd multiple of 180) out-of-phase with respect to light reflected at the ARC's upper surface (the interface between the DARC and the photoresist layer). Preferably, the intensity of the two reflections are similar, to maximize interference (i.e., cancellation). Reflections from other interlayer interfaces (both above and below the DARC) may also contribute to the radiant energy reflected, and may thus need to be accounted for when optimizing ARC characteristics. A process for forming a DARC should therefore provide accurate control over the rate at which the DARC is deposited.
In the prior art, thickness control is provided by the introduction of an inert gas, such as helium or argon, into the substrate processing chamber during the deposition of a DARC. Introducing such gases into the processing chamber at relatively high rates allows the flow rates of reactant gases to be reduced. These reduced flow rates translate into a reduction in the amount of material available for deposition during a given period of time, thereby reducing the film's deposition rate. However, gases such as helium and argon are relatively expensive and are not universally available.
Accordingly, improved methods for controlling the deposition rate of a film, thereby providing finer control over the film's thickness, are constantly being sought. Preferably, such an improved method should not rely on the use of substances that may be expensive or difficult to obtain. Additionally, it is preferable that such a method allow the control of other film parameters, such as the film's refractive and absorptive indices.